The 80/20 Engine: Where Your Profit Actually Lives

Bill Canady: Second floor conference room, power band parts, $310 million of revenue, 32% gross margin, 8% EBITDA, seven functional heads around the table, all of them competent, all of them tired, all of them holding a binder they spent the weekend preparing that I had not asked for. I asked him one question, not about their binders. I said, What's the percentage of your function's productive time? In a typical week is consumed by activities related to the bottom 20% of our customers. It's total silence. Then the kind of shifting that happens when people realize the question is real. Engineering went first. 19%, special drawings, one-off tolerances, quotes that never convert. Finance said 22. Collections, credit exceptions, disputed invoices, the manual work of billing people who pay in 91 days. Wow. That takes a long time. Generally, remember, it only comes after thirty. So operation scheduling said twenty-eight, expedites, changeovers, short runs. And ahead of customer said forty one percent. Not apologetically. She said it in a way you say a number you've been carrying around for a while and you just hope somebody will ask. The lowest answer in that room that day was 18%. Highest was forty one. Average across every function was about twenty seven percent. Then the CFO, the good CFO, quiet the whole time, looked up from his laptop and said, Those customers are about five percent of revenue. twenty seven percent of the company dedicated to deliver five percent of the revenue. Nobody in that room was underperforming. They were working extremely hard on the wrong third of the business, and not one report they received every month would ever have told them so. So here's the proof and the promise and the plan to deal with it. Starting with the proof. In a typical middle market business, the top 20 to 25% of your customers produce between 105 and 150%. You heard that right. More than the total piece of your profit. Read it again. It's more than 100%. That is not a typo and it is not enthusiasm. It's the most important sentence in this whole episode. Because the only way a portion of this business can produce more than all of your profit is if the rest of the business is producing negative profit. Your best customers aren't carrying the custom the company. They're carrying the company and subsidizing the part of it that's losing money. And the blended average on your P and L hides both facts at once. At the other end, the bottom 20% of your customers typically account for about 5% of revenue. About 30% of capacity and about 60% of complaints. Now here's my promise. By the end of the next half hour, you'll be able to take every dollar of revenue in your company and put it in one of four boxes. Know exactly what action belongs to each box, and ask one question in your next leadership meeting that will tell you in about 11 minutes how much your company is currently occupied by customers who aren't paying for it. Here's the plan. It's three moves. The first one is a data pull you can request today. And the second one is the analysis. And the third one is a question you ask out loud in a room. I'll tell you how each one's done, and I'll tell you the three specific ways people run this analysis and get a useless answer because that's more common than getting no answer at all. This is a thousand day CEO. I'm Bill Canady Let's get into it. Let's start with the problem. 80 twenty has been ruined by being repeated. You hear it all the time. Everybody's heard it, everybody nods, and most people think it means focus on what matters, which is not a principle, it's a greeting card. You cannot act on it. Nobody has ever walked out of a meeting focused on what matters and gained a point of margin. So let me restate it in exactly what it is. It's a diagnostic, it's not a slogan, it's not a philosophy, it's not even a mindset. It's a measurement you run on your own transaction data that tells you which specific customers are buying and which specific profits generate all the profit in your business and which ones consume it. The critical 20% that drives 80% of the results, the vital few versus the trivial many. And when you actually run it on real data at a real company, the ratio is usually saw sharper than 80-20. In practice, In the business I've done this in, it lands closer to 15 or 20% of inputs, producing 80% of the profit. That's right, even a smaller amount. So why doesn't the CEO already know this? Because your PL is an average and it's organized around the wrong thing. Look at how your financials get cut by legal entity, by plant, by region, by function, sales costs here, engineering costs there, operations down at the bottom. Every one of those cuts is useful for something. Not one of them shows you profit. Because profit doesn't live in a function or a plant or a region. Profit lives at the intersection of specific customers buying a specific product. That's the atomic unit. That is where almost certainly no report on your desk in any month shows you or your business cut that way. Which means the number you're managing to. 32% gross margin, 8% EBITDA is a blend. And the blend has two entirely different businesses inside of it. One of them is excellent. Concentrated customers, core products, high margin, low touch, predictable. If you owned only that business, you'd be at 15 to 18% EBITDA, and the sponsor would be delighted. So would you. And the other one is a business that loses money on nearly every transaction and is bolted to the first one. And it is consuming the first one's capacity to do it. Average them together and you get 8% EBADA and a board asking reasonably, why isn't this 12%? Now the cost of not knowing it isn't just the money you lose on the bad transaction, although that's real. The larger cost is that you allocate management attention evenly across a business where profit is distributed wildly unevenly. Your best customers, the ones producing more than 100% of your profit, get a quarterly check-in and a Christmas card. Your worst customers get more service than that because they complain more and the complaints route to attention. You have to build, without deciding to, a system that spends the most on the accounts that pay the least. You've just come out of meeting one with a goal, meeting two, set the strategy and Is due in weeks three through six, and strategy is a decision about where you will and won't compete. You cannot make that decision, honestly, without that analysis in your hand. Do it in the wrong order, and you will set a strategy on top of the same averages that got you here. And that's a version of this that costs you even when nothing goes wrong. Two customers, same revenue, same product, same price. One orders twelve times a year in full palace and pays in 30 days. We like that. The other orders 140 times a year in singles, calls before each one, changes half of them after their release, and pays in 80. Whew, we don't like that. Your PL changes both of them at the same cost of goods and the same allocated overhead because that's what allocation does. It spreads costs by revenue, which is precisely the assumption the whole exercise is designed to break. One more reason people skip it, and I'll say this plainly. The answer is uncomfortable. it frequently implicates a logo the company is proud of. And once in a while it implicates a customer the CEO personally brought in. That's the real reason the analysis doesn't get run, and it isn't a data problem. On two axes. On the first axis, your customers rank by value, not by revenue, by profit contribution. Split them, the customer A, the A customers, roughly the top quartile or two, and then the ones producing the 105 to 150%. And then next to B customers, everyone else. On the other axis, your products. And they're sorted by strategic fit. Core products, what you've actually built to make, where they have cost advantage, where the volume runs, the tail products, the accumulated inventory of everything you've ever agreed to make for anyone, for some reason, once. Cross them, and every dollar revenue in your company falls in exactly four boxes. Before I name the boxes, run the profit curve once, because seeing its shape is what makes the boxes obvious. Now take your customers, sorted best to worst, on profit, and add them up cumulatively. The line climbs steeply through the first quartile. And by the time you're through the top 20 or 25% of accounts, you've already above 100% of the profit that the company earned that year. That's where it starts standing out. Then the loan line goes flat for a long stretch. That's the middle of your book. The accounts that essentially cover their own costs. And nothing else. And then in the last stretch, it turns down and gives back somewhere between five and thirty percent of what you made. That's the negative. The downturn at the end is not a rounding era and it's not a data problem. It's a portion of your business that you are paying to keep. Every business I've run this on has that slope. Yours will too. The four quads. Quad one, the core, the fort. A customers buying core products. This is where nearly all the profit lives. Not most, nearly all. In some businesses I've measured, quad one is under quarter of the revenue and more than the entirety of the profit. The action here is protect and grow. And I want to slow down on protect because it's a word people skip. Quad one is where your most competitors, best salespeople spend their time. It is also almost universally the most underserved part of the business because these accounts don't complain. They're easy. They pay on time, and easy customers are invisible in a management system that routes attention towards noise. Ask yourself when your top ten profit accounts last got an unprompted visit from anyone senior. When did you go? In most companies, the honest answer is they didn't, and the account screaming at you last week did. Quad two, develop a customers buying tell products. These are good customers. They're proven. They're credit worthy. They're already buying from you, who are buying the wrong things from you. The action here is cross-sell the core. This is the cheapest growth available to any market business, and almost nobody works it systematically. Because sales compensation pays the same for a dollar of tail as a dollar of core. Now, Quad 2 is a list of names you already have with a relationship you already own. And the entire move is changing the mix of what they buy. Get them to buy the A products. Quad three, this is the one you earn in. B customers buying core products. They're buying the right things. They're just small or new or lower margin because they were priced on the way in. Or maybe they're just a small customer. The action is graduate or hold. Some of these are next year's A customers and should be worked deliberately towards that. The rest are fine as they are. Serve them efficiently at standard terms, at a standard price, with no special handling, and don't spend management's attention trying to make them into something they are. This is where automation comes in. Think ordering online. Quad four, the TEL. Be customers buying TEL products. Small, low value customers buying small, low value, non core products. This is the complexity tax. This is where profits go to die. The action is reprice or exit. And I'm going to say the next part carefully because it is a single point in this episode where people flinch. Firing or reprice in Quad 4 feels like losing revenue. Well, it's not. That revenue is consuming a third of your capacity and most of the complaints that you get to earn almost nothing. In fact, a lot of times it's a negative. Somebody in your room is going to say, but there's still $4 million of revenue. And they'll be right about the $4 million, but they'll be wrong about what it is. That 4% isn't the cost. The 4% is the bait. It's what keeps you paying for a cost structure, size for a company you don't want to be. Let's start with the capacity diagnostic. Now, here's the question: if you take one thing out of this episode, take this because it costs nothing. It takes 11 minutes and it works before you have a single line of data. In your next leadership meeting, go around the table, one function head at a time, and ask each one of them in these words. What percentage of your function's productive time in a typical week is consumed by activities related to the bottom twenty percent of our customers? Then do two things. Make them answer that with a number, not a sentence, and write down what each one says, with their name next to it in front of them. At Powerband Parse the answer were these operations scheduling twenty eight percent. Finance twenty two. Engineering nineteen. The lowest answer in the room was 18. The highest was 41. The average across every function was about 28%. And the sales number was its own conversation. Roughly a third of Phil's sales time, the most expensive selling hours in our company, was going to accounts that produce almost nothing. Not because those salespeople were lazy, because small accounts need more help per order, and territories were built by geography rather than value. And every one of those reps were doing exactly what the system asked. 27% of $310 million company for 5% of the revenue. Lots of effort, little reward. Now say someone always says the estimates are solved. Well they are. These are guesses, and they'll be off by a few percentage points either way. Doesn't matter. Nobody guesses twenty-seven when the truth is four. The direction and the magnitude of what you're after here, and you got them in eleven minutes for free. And there's a second thing that happens in that room that's worth much as the number. Once the head of engineering says nineteen percent out loud in front of his peers, with the CEO writing it down, the analysis you're about to run stops being an attack on the sales organization. It's now everyone's number. They say it to themselves. You're not going to have to convince that room of anything or In three weeks from now. Occupied means not available. It doesn't mean the same. One last piece, and this is the one that determines where any of these turn in the money. Say your VP of quality tells you she spends 30 minutes of her week on bottom quartile accounts. The instinct is to think. That's great. Excellent. I have 30% of a quality VP available to redeploy. You don't. She is not 30% available. She's 30% occupied. This is not a word game. It's the difference between a plan that works and a plan that produces exhaustion. Occupied capacity has to be released by a decision before it can be reallocated. Someone has to raise a price, set a more minimum order value, move an account to a distributor, or say no to a customer until one of those decisions is made and holds. The 30% is spoken for, and any new initiative you put on her plate is going on top of it. This is where most 8020 programs die and they die quietly. The analyst is right, the deck is good, the new growth initiatives get assigned, and nothing gets removed. Six months later, the leadership team is doing everything they were doing before, plus a growth agenda, and the CEO cannot understand why the team is strong, is moving away and This slowly. Why's nothing happening here? You have to free capacity before you can reallocate it. Free it first, then reallocate in that order every time. I know you just want to tell them to do it, but you gotta free them up. Now the plan. It's three moves. The first one you can request before lunch. Move number one: pull the 12-month transaction file. One file, four columns minimum, customer product, revenue, margin. Line item three, 12 months, so seasonality doesn't distort it. Ask your CFO for it directly and give the specification out loud because the request will get interpreted otherwise. What you want is the raw transaction extract. You not the summary. You need the raw data here. And here's the instruction that matters most. And you have to give it explicitly, or you will not get it. Do not Clean it. They to give it to you where it's all prettied up. Don't let them do it. Give you the raw data. The weird stuff is the finding. A file that has been scrubbed clean in a tie into tidy shape is a file that has had the answer scrubbed out of it. Tell them raw extract. 12 months, four columns, one file this week. You'd rather have it ugly on Friday than beautiful in a month. Second move, build the quad map. Rank customers by profit contribution and split A from B. Sort products into a cortel and let the ed operator make the call, not a spreadsheet. Because core is a judgment about what your business is built to do, not a percentile. Then put every transaction in its box and total the four boxes on two dimensions: percentage of revenue and percentage of profit. That's the deliverable. Four boxes, two numbers each. It fits on an index card and it will be the most valuable index card in your building. Now, there's three warnings here because I've watched all three of these ruin the work. First, don't do it at an account name level. If you grade whole customers and so transactions, you'll lose Quad 2 entirely. And Quad 2 is the cheapest growth you own. Big customers buy tail products too. Small customers buy core. The analysis only works at the transaction level, which is precisely why the account name shortcut is so tempting. It's much faster and it gives you an answer that feels right, but it is wrong. Second, don't let sales grade their own customers. You cannot ask the lion to guard the antelope. Every account will come back strategic. Everyone will have a reason. Long relationship growing next year gets us into a vertical. The parent company is enormous. Grade the accounts on the data first in a room without the account owners in it. Then bring sales in to challenge the specific cases with evidence. The order matters. Data first, argument second, and the argument has to move the number to win. Number three, don't stop at the chart. This is the most common failure and the most expensive one. The analysis gets built. It's genuinely good. It gets presented to the leadership team. Everybody agrees it's illuminated. And then it becomes a document. Six weeks later it's in a folder and the business is running exactly as it was. The chart's not the deliverable. The decision is. Nothing has happened until a price has moved, an account has been returned, or a product has been retired retired. What it looks like when it's working and when it's theater. When it's working, the language in the building changes from and you find out about it secondhand. Nobody is a pricing conversation. Somebody in a pricing conversation says, that's a quad four order. And everybody in the room knows what that means and what happens next. The quad map gets updated quarterly without you asking. And your top 10 profit accounts start appearing on senior calendars on purpose, unprompted. When it's theater, the analysis is impeccable. And nothing moves. There's a steering committee, there's a phase two. The word segmentation appears in three decks, and this revenue is exactly where it is, which is what people describe as encouraging. And here's the tell I use. 90 days after the analysis lands, ask one question: What revenue have we walked away from? If the answer is none, you didn't do the work, you did the reading. Okay, the number. Three things to keep from this one. Number one, the top 20 to 25% of your customers produce a hundred and five to a hundred and fifty percent of your profit. And the reason it exceeds a hundred is that the bottom is negative. They're taking it away. Your PL is an average, and the average is hiding two businesses that need completely different management systems. Number two, four quads, quad one, the core, the fort, protect and grow. Nearly all the profit lives there. Quad two, Cross-sell the core and it's the cheapest growth you own. Quad three, graduate or hold. Quad four, reprice or exit. Firing or repricing quad four feels like losing revenue. It's not. That revenue is consuming a third of your capacity and most of your complaints to earn almost nothing. Number three, occupied is not available. This is a good one and it's tough to handle. A manager spends 30% of her week on a bottom quartile account. Is not 30% available to you. She's 30% occupied, and that capacity does not come free until somebody makes a decision that releases it. Free first, then reallocate. That's your job. And the one question which costs you nothing in 11 minutes: what percentage of your function's productive time in a typical week is consumed by activities related to the bottom 20% of your customers? Ask it this week. Write the numbers down. When the lowest answer in your room comes back at 18%, you'll understand why your leadership team is exhausted and your EBA dies flat at the same time. And those two fit facts will finally be the same fact. Luck is the residual of design, and this is where you find out what you design. Next week is the one I'd pick if you only listened to a single episode in this season. It's a ratio. Two lines off your P and L, about thirty minutes of arithmetic, and it returns one verdict. Whether your company has earned the right to grow at all, and whether you're right now would make your business bigger, busier, and poorer. There's a threshold number, and I'll give it to you, and the most middle market businesses I've measured come in under it the first time. You want the quad template and the capacity diagnostic, all the questions and the tables to write them down in. They're all available at the eighty twenty institute dot com. Lots of courses and lessons there for you to come in and try them out. It's a diagnostic, and these things typically take about thirty minutes. I'm Bill Canady and I'll see you at the next hundred days.